Releases: davidusb-geek/emhass-add-on
Releases · davidusb-geek/emhass-add-on
EMHASS add-on v0.14.0
0.14.0 - 2025-11-15
Improvement
- Add InfluxDB integration for enhanced historical data retrieval (@scrusberghs)
- Added SSL feature for InfluxDB (@sokorn)
- Feature to export influxdb to csv file (@scrusberghs)
- Improvements to timezone handling and testing, use ISO8601 date format (@Crosenhain)
- Reduce severity of logged Open Meteo
forecast_daysminimum override message (@paulhomes)
Fix
- Fix to InfluxDB query to avoid TypeError: Cannot compare tz-naive and tz-aware timestamps (@Jahper)
- Fix InfluxDB future data query and
use_influxdbparameter type (@scrusberghs) - Resolve time zone handling issues across daylight savings changes (@Squazel)
- Fix
forecast.pyto align CSV data with prediction window (@nielskool) - Fix for
pytz.exceptions.NonExistentTimeErrorwith Open Meteo weather source during DST change (@paulhomes) - Fix decimal precision for
unit_load_costandunit_prod_pricesensors (@scrusberghs)
Full Changelog: v0.13.5...v0.14.0
EMHASS add-on v0.13.5
Improvement
- Use SVG render mode for Plotly figures (@Crosenhain)
- Constrain PV curtailment of hybrid inverter (@rmounce)
Fix
- Fix operating_timesteps_of_each_deferrable_load validation formula (@scruysberghs)
Full Changelog: v0.13.4...v0.13.5
EMHASS add-on v0.13.4
A new release of EMHASS with a bunch of fixes and new features...
Improvement
- Added support for minimum power value for each deferrable load
- Add configuration for hybrid inverter input power limit and efficiency (@rmounce)
- Add runtime parameter to prevent PV forecast feedback loop during curtailment (@scruysberghs)
- Update Docker documentation for volume mounting of config.json (@cabberley)
- Apply ruff formatting for improved code consistency (@polyfloyd)
- Set executable permissions for
__main__entrypoints and remove unused shebangs (@polyfloyd) - Correct minor typos in README documentation (@Ultimation)
Fix
- Fixed naive load power forecast shifted timestamp issue #516
- Fix Docker build for aarch64 architecture (@polyfloyd)
EMHASS add-on v0.13.3
Improvement
- Enhance optimization solver capabilities by adding HiGHS solver, improving thread configuration, and updating solver support across the project (@Crosenhain)
- Update spelling and extended some error messages to include more troubleshooting information (@kcoffau)
Fix
- Fixed incorrect indentation in a for loop, leaving relevant constraints out of the optimisation (@purcell_lab)
Full Changelog: v0.13.2...v0.13.3
EMHASS add-on v0.13.2
Improvement
- Improved README and development documentation (@Crosenhain and @gieljnssns)
- Improved forecasting and optimization classes logging (@purcell_lab)
Fix
- Fixed ML Regressor for correct time zone treatment (@gieljnssns)
- Fixed long standing problem to allow thermal and standard loads to be scheduled (@purcell_lab)
- Fixing Missing files in python package issue (@pail23)
- Fixed improper handling of home assistant API errors (@Gruek)
- Remove an extraneous comma in the pd.to_datetime() method call in retrieve_hass.py (@GeoDerp)
- Fixed the SettingWithCopyWarning Pandas problem and PV sensor handling when set_use_adjusted_pv is false (@paulhomes)
Full Changelog: v0.13.1...v0.13.2
EMHASS add-on v0.13.1
0.13.1 - 2025-04-21
Improvement
- Add timeout and thread configuration for different LP solvers (CBC, GLPK, COIN). Thanks to @GeoDerp
- Clean up actions and add sourcery scan action. Thanks to @GeoDerp
- Cache Open-Meteo JSON reponse locally to reduce the number of API calls and add resilience. Thanks to @paulhomes
Fix
- Update default weight battery value in config_defaults.json. Thanks to @purcell-lab
- Fix forecast output assignment and avoid SettingWithCopyWarning. Thanks to @radim2tej
- Fix for get_mix_forecast ValueError: cannot convert float NaN to integer. Thanks to @paulhomes
- Fix for hard-coded actual/forecast PV sensor names. Thanks to @paulhomes
- Update pyproject.toml to solve issue #495: Missing files in python package after 0.12.5. Thanks to @GeoDerp
Full Changelog: v0.13.0...v0.13.1
EMHASS add-on v0.13.0
A new method to adjust PV forecast!
Checkout the upgraded documentation here: https://emhass.readthedocs.io/en/latest/forecasts.html#adjusting-pv-forecasts-using-machine-learning
Improvement
- Added a new feature to automatically adjust PV forecast using historical actual/forecast data
- Added Support for multiple days forecasting with CSV data. Thanks to @nielskool for this contribution!
- Refactored a new method
retrieve_home_assistant_dataincommand_line - Unified default training data for ML algorithms using
data/long_train_data.pkl - Added a presentation stack using
marp
Fix
- Fixed Value has dtype incompatible with int64 error, issue #465
- Fixed Runtime "delta_forecast_daily" not recognized, issue #479
- Attempt to fix Missing files in python package after 0.12.5, issue #495
- Fixed workflow for macos tests
Full Changelog: v0.12.8...v0.13.0
EMHASS add-on v0.12.8
0.12.8 - 2025-02-27
Improvement
- Added a new method for weather forecast: open-meteo. This offers a very nice API to obtain weather forecasts. We are dropping the old scrapper method, if you were using the scapper method you just need to change the Weather forecast method to open-meteo
- Added support to
device_classdefinition when posting sensors to HA. Thanks to @paulhomes for this contribution - Improved github workflow Python testing
- Updated documentation
Fix
- Fixed TZ shift issue when using the typical load power forecaster
Full Changelog: v0.12.7...v0.12.8
EMHASS add-on v0.12.7
0.12.7 - 2025-02-16
Improvement
- A first batch of improvements (code reliability) proposed by SonarQube
Fix
- Added missing files to Dockerfile
- Changes to the Dockerfile to make
/app/datafolder persistent. Thanks to @scruysberghs - Fixing errors with missing sensor
power_load_no_var_loads. Thanks to @scruysberghs - Improved documentation and fixed automated testing for code quality. Thanks to @GeoDerp
Full Changelog: v0.12.6...v0.12.7
EMHASS add-on v0.12.6
0.12.6 - 2025-02-09
Improvement
- Improved warning message in case of wrong URL setting
- Improved testing on typical load forecast method with different time steps
Fix
- Solved missing file problem: data_load_cost_forecast.csv
- Fixed warning issue on
findallmethod
Full Changelog: v0.12.5...v0.12.6